EUVIMEDScopri EUVIMED
Open access · CC BY 4.0

Super‐Resolution Enhanced Defect Detection Network for Photovoltaic Panel Images

Hao Tang, Bin Yang

IET Image Processing · 2026-01 · 20 (1)

DOI: 10.1049/ipr2.70436 ↗

At a glance

Journal article in IET Image Processing, published online on 2026-07-20. Written by Hao Tang and 1 co-authors. 40 references. The full text is freely available (open access, CC BY 4.0). Funded by: Natural Science Foundation of Shanghai Municipality.

AbstractDati bibliograficiCiteRelated studies

Abstract (original language)

Reliable defect detection of photovoltaic (PV) modules is vital for maintaining energy efficiency and ensuring long‐term system stability. To overcome typical challenges such as low‐resolution inputs, high inspection costs and frequent omission of minor defects, a super‐resolution (SR) enhanced method, PVSR‐YOLO, is proposed to improve the original YOLOv10s architecture for PV defect detection. A novel SR feature learning branch is introduced to reconstruct high‐resolution images from low‐resolution inputs, guiding the backbone to learn fine‐grained defect features during training. The SR branch is discarded during inference, enabling improvements in accuracy without increasing computational cost. In addition, another SR augmentation network is applied to preprocess the input dataset, generating high‐resolution images for training the SR branch. Experiments on two PV defect datasets demonstrate that the proposed method outperforms baseline YOLO models and state‐of‐the‐art PV defect detectors. It achieves 86.1% mAP50 on the PV Multi‐Defect dataset, with remarkable improvements in small‐defect categories, verifying its robustness and promising reliability for PV defect detection.

Funding

  • Natural Science Foundation of Shanghai Municipality

Cite

APA

Tang, H., & Yang, B. (2026). Super‐Resolution Enhanced Defect Detection Network for Photovoltaic Panel Images. IET Image Processing, 20(1). https://doi.org/10.1049/ipr2.70436

Vancouver

Tang H, Yang B. Super‐Resolution Enhanced Defect Detection Network for Photovoltaic Panel Images. IET Image Processing. 2026;20(1). doi:10.1049/ipr2.70436

BibTeX
@article{Tang2026,
  author = {Tang, Hao and Yang, Bin},
  title = {Super‐Resolution Enhanced Defect Detection Network for Photovoltaic Panel Images},
  journal = {IET Image Processing},
  year = {2026},
  volume = {20},
  number = {1},
  doi = {10.1049/ipr2.70436}
}
Start for free →